2011/1/6 Simon Ihmig <ih...@kaliber5.de> > Hi there, > > I am absolutely new to pythonocc. I just want to ask if it might suit the > task I am trying to solve, before digging too much into it... > > So here we go: > > There is an existing application, falling into the category of "product > configurators", that needs to be extended to suppport a 3D view of the > configurated product (it's just 2D untill now). One component of that task > is generating the 3D geometry dynamically. That's where I thought pythonocc > might come into play. > > So the component should preferably be used as a command line programm, > accepting the product configuration as input and outputting a file with the > corresponding 3d geometry, preferably in Collada Format. > > The needed 3d operations are: > - Adding geometry from existing files (no big deal) > - Transforming that geometry (translation, rotation, scaling) (no big deal > either I assume) > - creating new geometry by sweeping a surface, in this case by rotating > around the y axis (lathe tool) > - performing boolean operations (cut) > > The output should be a relatively low poligon count mesh, suitable for fast > display with OpenGL on standard PC hardware (no fancy graphics hardware), > with different identifiable surface materials (the actual shading can be > definied later, as soon as the tool has done the job of outputting the > collada file. However the different parts of geometry with different > material types, e.g. metal, glass etc., must be separatable). The file > should be as I said in collada format, which is not available out of the box > form pythonocc as I understand, so maybe some conversion from vrml to > collada is necessary? > > The whole process should not take substantialy more than a second, > preferably less, so it can be integrated in an interactive application > without delaying the user experience. > > So can this be done with pythonocc, hopefully in an efficient and easy > manner? > > Thanks a lot for your advice! > > Simon >
Hi Simon, I would answer 'yes' to all of your questions : pythonocc is suitable for your project *but* the criteria 'less than 1s' to perform the different operations has to be considered regarding the type of product. If the configurable product is a car or an aircraft, with thousands of parts, then it's not possible to achieve configuration (from input to 3D output) in less than 1s. Actually, I doubt that any CAD kernel would do it, even with mutiprocessed/multitasks kernels. I didn't benchmarked pythonocc in terms of computation speed, and I can't be more precise. But such a dataset would be interesting: a truly industrial solution has to be stressed with huge models. Usual live demos of commercial products only deal with 'small products', but many issues only occur when the model size becomes bigger : whereas vendors do not communicate about the weaknesses of their product in such cases, pythonOCC should be highly stressed and results made publicly available. Regards, Thomas >
_______________________________________________ Pythonocc-users mailing list Pythonocc-users@gna.org https://mail.gna.org/listinfo/pythonocc-users
